
import pandas as pd
import numpy as np
from scipy.stats import zscore

import seaborn as sns
import plotly.express as px
import matplotlib.pyplot as plt
import plotly.graph_objects as go
from plotly.subplots import make_subplots

from yellowbrick.cluster import silhouette_visualizer

import datetime

from sklearn import preprocessing

from scipy.stats import zscore

from dataloader import DataLoader

from clustering import KMeansCluster, CMeansCluster

from visualization import tsne

from pathlib import Path

images_path = Path("./images2")

np.random.seed(42)


df = pd.read_csv('energy-usage-2010-clean.csv')
df


df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 67051 entries, 0 to 67050
Data columns (total 73 columns):
 #   Column                               Non-Null Count  Dtype  
---  ------                               --------------  -----  
 0   COMMUNITY AREA NAME                  67051 non-null  object 
 1   CENSUS BLOCK                         66974 non-null  float64
 2   BUILDING TYPE                        66974 non-null  object 
 3   BUILDING_SUBTYPE                     66974 non-null  object 
 4   KWH JANUARY 2010                     66180 non-null  float64
 5   KWH FEBRUARY 2010                    66180 non-null  float64
 6   KWH MARCH 2010                       66180 non-null  float64
 7   KWH APRIL 2010                       66180 non-null  float64
 8   KWH MAY 2010                         66180 non-null  float64
 9   KWH JUNE 2010                        66180 non-null  float64
 10  KWH JULY 2010                        66180 non-null  float64
 11  KWH AUGUST 2010                      66180 non-null  float64
 12  KWH SEPTEMBER 2010                   66180 non-null  float64
 13  KWH OCTOBER 2010                     66180 non-null  float64
 14  KWH NOVEMBER 2010                    66180 non-null  float64
 15  KWH DECEMBER 2010                    66180 non-null  float64
 16  TOTAL KWH                            66180 non-null  float64
 17  ELECTRICITY ACCOUNTS                 66180 non-null  float64
 18  ZERO KWH ACCOUNTS                    67051 non-null  int64  
 19  THERM JANUARY 2010                   64821 non-null  float64
 20  THERM FEBRUARY 2010                  62819 non-null  float64
 21  THERM MARCH 2010                     65569 non-null  float64
 22  THERM APRIL 2010                     65476 non-null  float64
 23  THERM MAY 2010                       65194 non-null  float64
 24  THERM JUNE 2010                      65284 non-null  float64
 25  THERM JULY 2010                      65231 non-null  float64
 26  THERM AUGUST 2010                    65143 non-null  float64
 27  THERM SEPTEMBER 2010                 64769 non-null  float64
 28  THERM OCTOBER 2010                   65329 non-null  float64
 29  THERM NOVEMBER 2010                  65492 non-null  float64
 30  THERM DECEMBER 2010                  65507 non-null  float64
 31  TOTAL THERMS                         65755 non-null  float64
 32  GAS ACCOUNTS                         65755 non-null  float64
 33  KWH TOTAL SQFT                       65901 non-null  float64
 34  THERMS TOTAL SQFT                    65378 non-null  float64
 35  KWH MEAN 2010                        66180 non-null  float64
 36  KWH STANDARD DEVIATION 2010          57095 non-null  float64
 37  KWH MINIMUM 2010                     66180 non-null  float64
 38  KWH 1ST QUARTILE 2010                66180 non-null  float64
 39  KWH 2ND QUARTILE 2010                66180 non-null  float64
 40  KWH 3RD QUARTILE 2010                66180 non-null  float64
 41  KWH MAXIMUM 2010                     66180 non-null  float64
 42  KWH SQFT MEAN 2010                   65901 non-null  float64
 43  KWH SQFT STANDARD DEVIATION 2010     51666 non-null  float64
 44  KWH SQFT MINIMUM 2010                65901 non-null  float64
 45  KWH SQFT 1ST QUARTILE 2010           65901 non-null  float64
 46  KWH SQFT 2ND QUARTILE 2010           65901 non-null  float64
 47  KWH SQFT 3RD QUARTILE 2010           65901 non-null  float64
 48  KWH SQFT MAXIMUM 2010                65901 non-null  float64
 49  THERM MEAN 2010                      65755 non-null  float64
 50  THERM STANDARD DEVIATION 2010        56821 non-null  float64
 51  THERM MINIMUM 2010                   65755 non-null  float64
 52  THERM 1ST QUARTILE 2010              65755 non-null  float64
 53  THERM 2ND QUARTILE 2010              65755 non-null  float64
 54  THERM 3RD QUARTILE 2010              65755 non-null  float64
 55  THERM MAXIMUM 2010                   65755 non-null  float64
 56  THERMS SQFT MEAN 2010                65378 non-null  float64
 57  THERMS SQFT STANDARD DEVIATION 2010  51367 non-null  float64
 58  THERMS SQFT MINIMUM 2010             65378 non-null  float64
 59  THERMS SQFT 1ST QUARTILE 2010        65378 non-null  float64
 60  THERMS SQFT 2ND QUARTILE 2010        65378 non-null  float64
 61  THERMS SQFT 3RD QUARTILE 2010        65378 non-null  float64
 62  THERMS SQFT MAXIMUM 2010             65378 non-null  float64
 63  TOTAL POPULATION                     67037 non-null  float64
 64  TOTAL UNITS                          67037 non-null  float64
 65  AVERAGE STORIES                      67051 non-null  float64
 66  AVERAGE BUILDING AGE                 67051 non-null  float64
 67  AVERAGE HOUSESIZE                    67037 non-null  float64
 68  OCCUPIED UNITS                       67037 non-null  float64
 69  OCCUPIED UNITS PERCENTAGE            64606 non-null  float64
 70  RENTER-OCCUPIED HOUSING UNITS        67037 non-null  float64
 71  RENTER-OCCUPIED HOUSING PERCENTAGE   64433 non-null  float64
 72  OCCUPIED HOUSING UNITS               67037 non-null  float64
dtypes: float64(69), int64(1), object(3)
memory usage: 37.3+ MB


dl = DataLoader(df)


all_energy_df = dl[dl.energy_cols]
all_gas_df = dl[dl.gas_cols]


figure = plt.figure(figsize = (20,10))

ax = sns.boxplot(data=all_energy_df)
_ = ax.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=30)
_ = ax.set(xlabel="Mes", ylabel="KWH")

path = images_path / Path("exploracion_datos/outliers/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("boxplot.eps"))


figure = plt.figure(figsize = (20,10))

energy_all_mean = all_energy_df.mean(axis=0)
energy_all_sd = all_energy_df.std(axis=0)
ax = sns.barplot(data=energy_all_mean, x = energy_all_mean.index, y = energy_all_mean.values, yerr=energy_all_sd)
_ = ax.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=30)
_ = ax.set(xlabel="Mes", ylabel="KWH de media")

path = images_path / Path("exploracion_datos/outliers/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("barplot.eps"))


figure = plt.figure(figsize = (20,10))

energy_all_mean = all_energy_df.mean(axis=0)
ax = sns.barplot(data=energy_all_mean, x = energy_all_mean.index, y = energy_all_mean.values)
_ = ax.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=30)
_ = ax.set(xlabel="Mes", ylabel="KWH de media")

path = images_path / Path("exploracion_datos/outliers/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("barras.eps"))


figure = plt.figure(figsize = (20,10))

ax = sns.boxplot(data=all_gas_df)
_ = ax.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=30)
_ = ax.set(xlabel="Mes", ylabel="Therms")

path = images_path / Path("exploracion_datos/outliers/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("boxplot.svg"))


figure = plt.figure(figsize = (20,10))

gas_all_mean = all_gas_df.mean(axis=0)
gas_all_sd = all_gas_df.std(axis=0)
ax = sns.barplot(data=energy_all_mean, x = gas_all_mean.index, y = gas_all_mean.values, yerr=gas_all_sd)
_ = ax.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=30)
_ = ax.set(xlabel="Mes", ylabel="Therms de media")

path = images_path / Path("exploracion_datos/outliers/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("barplot.svg"))


figure = plt.figure(figsize = (20,10))

gas_all_mean = all_gas_df.mean(axis=0)
ax = sns.barplot(data=gas_all_mean, x = gas_all_mean.index, y = gas_all_mean.values)
_ = ax.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=30)
_ = ax.set(xlabel="Mes", ylabel="Therms de media")

path = images_path / Path("exploracion_datos/outliers/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("barras.svg"))


dl['COMMUNITY AREA NAME'].drop_duplicates()

0          Archer Heights
1                 Ashburn
2          Auburn Gresham
3                  Austin
12               Avondale
              ...        
591           West Elsdon
602         South Deering
607    Washington Heights
736       Mount Greenwood
740       Near South Side
Name: COMMUNITY AREA NAME, Length: 77, dtype: object


energy_mean_by_community_type = dl[['COMMUNITY AREA NAME', 'BUILDING TYPE'] + dl.energy_cols].groupby(['COMMUNITY AREA NAME', 'BUILDING TYPE'], as_index=False).mean()
energy_mean_by_community_type


energy_commercial_by_community = energy_mean_by_community_type[energy_mean_by_community_type['BUILDING TYPE']=="Commercial"]
energy_commercial_by_community


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = energy_commercial_by_community[dl.energy_cols]
norm_data = scaler.fit(data).transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(4)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels, np.where(km_labels==0)[0]

(array([0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
        0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,
        0, 0, 0, 2, 2, 3, 0, 0, 0, 0, 0, 3, 0, 0, 3, 3, 0, 0, 0, 3, 0, 0,
        0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0]),
 array([ 0,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
        18, 19, 20, 21, 22, 23, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
        36, 37, 38, 39, 40, 42, 43, 44, 45, 46, 50, 51, 52, 53, 54, 56, 57,
        60, 61, 62, 64, 65, 66, 67, 68, 69, 70, 71, 73, 74, 75, 76],
       dtype=int64))


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


kmax = 10
k_values = [e for e in range(2, kmax)]

scaler = preprocessing.StandardScaler()
data = energy_commercial_by_community[dl.energy_cols]
norm_data = scaler.fit(data).transform(data)

cm = CMeansCluster(norm_data)
cm.cluster(k_values)

scores = cm.get_scores()
ax = sns.lineplot(data=scores, x="cluster_dim", y="score")
_ = ax.set(xlabel="Dimensión clúster", ylabel="Score")
_ = ax.set(xlim=(1.5, kmax + .5))

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow_cmeans.svg"))


cm_best = cm.get_cmeans_of(2)
cntr, u, u0, d, jm, p, fpc = cm_best

labels = u.argmax(axis=0)

labels, np.where(labels==1)[0]

(array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
        1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1,
        1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
        1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], dtype=int64),
 array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
        17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33,
        34, 35, 36, 37, 38, 39, 40, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,
        52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68,
        69, 70, 71, 72, 73, 74, 75, 76], dtype=int64))


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(10, 15), sharex=True)

for ki in range(k):
    cluster = energy_commercial_by_community.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    ax0.set(ylabel="kWh")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(energy_commercial_by_community[dl.energy_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {e}" for e in range(k)])

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in energy_commercial_by_community.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


cm_best = cm.get_cmeans_of(2)
cntr, u, u0, d, jm, p, fpc = cm_best

cntr = scaler.inverse_transform(cntr)
cm_labels = u.argmax(axis=0)

cm_labels

array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], dtype=int64)


labels = pd.DataFrame(cm_labels, columns=["label"])

k = cntr.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = energy_commercial_by_community.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=cntr[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = cntr[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    ax0.set(ylabel="kWh")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica_cmeans.svg"))


energy_residential_by_community = energy_mean_by_community_type[energy_mean_by_community_type['BUILDING TYPE']=="Residential"]
energy_residential_by_community


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = energy_residential_by_community[dl.energy_cols]
norm_data = scaler.fit(data).transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(4)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([3, 3, 3, 0, 3, 3, 3, 3, 3, 0, 3, 3, 3, 3, 3, 3, 0, 0, 0, 3, 3, 3,
       0, 3, 0, 3, 3, 0, 3, 3, 3, 3, 3, 3, 3, 3, 0, 3, 0, 3, 3, 1, 3, 3,
       3, 0, 0, 2, 2, 0, 3, 3, 3, 3, 0, 0, 3, 3, 3, 0, 3, 3, 3, 3, 3, 3,
       0, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = energy_residential_by_community.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    ax0.set(ylabel="kWh")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(energy_residential_by_community[dl.energy_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])
    
path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in energy_residential_by_community.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


energy_industrial_by_community = energy_mean_by_community_type[energy_mean_by_community_type['BUILDING TYPE']=="Industrial"]
energy_industrial_by_community


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = energy_industrial_by_community[dl.energy_cols]
norm_data = scaler.fit(data).transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(3)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 2, 0])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = energy_industrial_by_community.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    ax0.set(ylabel="kWh")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(energy_industrial_by_community[dl.energy_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_energia_tipo_zona_agrupa_ciudades/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in energy_industrial_by_community.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


gas_mean_by_community_type = dl[['COMMUNITY AREA NAME', 'BUILDING TYPE'] + dl.gas_cols].groupby(['COMMUNITY AREA NAME', 'BUILDING TYPE'], as_index=False).mean()
gas_mean_by_community_type


gas_commercial_by_community = gas_mean_by_community_type[gas_mean_by_community_type['BUILDING TYPE']=="Commercial"]
gas_commercial_by_community


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = gas_commercial_by_community[dl.gas_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(4)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 0, 1, 1, 1, 0,
       1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 3, 1, 1,
       1, 1, 1, 2, 0, 0, 0, 1, 1, 1, 1, 0, 1, 0, 2, 2, 1, 1, 1, 0, 1, 1,
       1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = gas_commercial_by_community.iloc[labels.query(f'label == {ki}').index][dl.gas_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} Therms")
    ax0.set(ylabel="Therms")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(gas_commercial_by_community[dl.gas_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in gas_commercial_by_community.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


gas_residential_by_community = gas_mean_by_community_type[gas_mean_by_community_type['BUILDING TYPE']=="Residential"]
gas_residential_by_community


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = gas_residential_by_community[dl.gas_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(4)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([3, 0, 3, 3, 3, 3, 3, 0, 0, 3, 0, 0, 0, 3, 3, 0, 0, 1, 3, 3, 0, 3,
       3, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 3, 3, 3, 0, 2, 0, 0,
       0, 3, 3, 3, 1, 0, 0, 0, 0, 3, 3, 1, 0, 3, 0, 3, 3, 0, 0, 0, 0, 3,
       3, 3, 3, 0, 0, 0, 3, 0, 3, 0, 0])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = gas_residential_by_community.iloc[labels.query(f'label == {ki}').index][dl.gas_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} Therms")
    ax0.set(ylabel="Therms")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(gas_residential_by_community[dl.gas_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in gas_residential_by_community.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


gas_industrial_by_community = gas_mean_by_community_type[gas_mean_by_community_type['BUILDING TYPE']=="Industrial"]
gas_industrial_by_community


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = gas_industrial_by_community[dl.gas_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(5)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 4, 3, 0, 0, 3, 0])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = gas_industrial_by_community.iloc[labels.query(f'label == {ki}').index][dl.gas_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} Therms")
    ax0.set(ylabel="Therms")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(gas_industrial_by_community[dl.gas_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter.svg"))


v = tsne(gas_industrial_by_community[dl.gas_cols], 3)

fig = go.Figure()
for label in range(k):
    fig.add_trace(
        go.Scatter3d(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1],
            z=v[(labels['label']==label).values, 2]
        )
    )
fig.show()

path = images_path / Path("ensayos/clustering_gas_tipo_zona_agrupa_ciudades/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter3d.svg"))

<Figure size 576x396 with 0 Axes>


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in energy_commercial_by_community.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


energy_df = dl[["COMMUNITY AREA NAME", "CENSUS BLOCK", "BUILDING TYPE", "BUILDING_SUBTYPE"] + dl.energy_cols]
communities = dl['COMMUNITY AREA NAME'].drop_duplicates()
communities

0          Archer Heights
1                 Ashburn
2          Auburn Gresham
3                  Austin
12               Avondale
              ...        
591           West Elsdon
602         South Deering
607    Washington Heights
736       Mount Greenwood
740       Near South Side
Name: COMMUNITY AREA NAME, Length: 77, dtype: object


commercial_loop = energy_df[(energy_df["COMMUNITY AREA NAME"] == "Loop") & (energy_df["BUILDING TYPE"] == "Commercial")]
commercial_loop


kmax = 7
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()

data = commercial_loop[dl.energy_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_energia_por_comunidad_observando_patrones/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(3)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 2, 0, 1, 0, 0, 0, 0, 0, 0, 1,
       0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 1,
       1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1,
       0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
       0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0,
       1, 0, 0, 0, 0, 1, 1, 1, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_energia_por_comunidad_observando_patrones/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = commercial_loop.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    ax0.set(ylabel="kWh")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_energia_por_comunidad_observando_patrones/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(commercial_loop[dl.energy_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_energia_por_comunidad_observando_patrones/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in commercial_loop.values[com_index][:,(0,1,3)]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


residential_loop = energy_df[(energy_df["COMMUNITY AREA NAME"] == "Loop") & (energy_df["BUILDING TYPE"] == "Residential")]
residential_loop


kmax = 5
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = residential_loop[dl.energy_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_energia_por_comunidad_observando_patrones/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

C:\Users\Alberto\miniconda3\envs\cdi-musiani-env\lib\site-packages\yellowbrick\utils\kneed.py:140: YellowbrickWarning:

No 'knee' or 'elbow point' detected This could be due to bad clustering, no actual clusters being formed etc.

C:\Users\Alberto\miniconda3\envs\cdi-musiani-env\lib\site-packages\yellowbrick\cluster\elbow.py:343: YellowbrickWarning:

No 'knee' or 'elbow' point detected, pass `locate_elbow=False` to remove the warning

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(3)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([1, 1, 2, 0, 1])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_energia_por_comunidad_observando_patrones/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = residential_loop.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    ax0.set(ylabel="kWh")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_energia_por_comunidad_observando_patrones/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(residential_loop[dl.energy_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_energia_por_comunidad_observando_patrones/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter.svg"))


v = tsne(residential_loop[dl.energy_cols], 3)

fig = go.Figure()
for label in range(k):
    fig.add_trace(
        go.Scatter3d(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1],
            z=v[(labels['label']==label).values, 2]
        )
    )
fig.show()

path = images_path / Path("ensayos/clustering_energia_por_comunidad_observando_patrones/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter3d.svg"))

<Figure size 576x396 with 0 Axes>


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in residential_loop.values[com_index][:,(0,1,3)]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


industrial_loop = energy_df[(energy_df["COMMUNITY AREA NAME"] == "Near West Side") & (energy_df["BUILDING TYPE"] == "Industrial")]
industrial_loop


f, axes = plt.subplots(3, 1, figsize=(5, 15), sharex=True)

for i, (census_num, (index, row)) in enumerate(zip(industrial_loop['CENSUS BLOCK'], industrial_loop[dl.energy_cols].iterrows())):
    ax0 = sns.lineplot(data=row.T, ax=axes[i], color="red")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    
    ax0.set_title(f"Census block {int(census_num)}")
    Media = row.mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    
    ax0.set(ylabel="kWh")

path = images_path / Path("ensayos/clustering_energia_por_comunidad_observando_patrones/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


gas_df = dl[["COMMUNITY AREA NAME", "CENSUS BLOCK", "BUILDING TYPE", "BUILDING_SUBTYPE"] + dl.gas_cols]
gas_df


commercial_loop = gas_df[(gas_df["COMMUNITY AREA NAME"] == "Loop") & (gas_df["BUILDING TYPE"] == "Commercial")]
commercial_loop


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = commercial_loop[dl.gas_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_gas_por_comunidad_observando_patrones/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(4)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([3, 3, 3, 3, 3, 3, 0, 0, 2, 3, 2, 3, 3, 3, 3, 3, 3, 3, 0, 0, 0, 3,
       1, 0, 3, 2, 0, 2, 3, 0, 0, 2, 2, 0, 3, 3, 3, 0, 3, 3, 0, 0, 3, 0,
       3, 3, 3, 3, 0, 3, 3, 3, 0, 3, 3, 3, 0, 3, 3, 0, 3, 3, 0, 3, 3, 3,
       3, 3, 3, 3, 3, 3, 3, 3, 0, 3, 3, 3, 3, 0, 3, 0, 3, 0, 3, 3, 3, 0,
       0, 3, 3, 0, 3, 3, 3, 0, 3, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 0, 0,
       3, 2, 3, 3, 3, 0, 3, 3, 3, 0, 3, 3, 3, 3, 3, 3, 3, 0, 3, 3, 3, 3])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_gas_por_comunidad_observando_patrones/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = commercial_loop.iloc[labels.query(f'label == {ki}').index][dl.gas_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} Therms")
    ax0.set(ylabel="Therms")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_gas_por_comunidad_observando_patrones/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(commercial_loop[dl.gas_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_gas_por_comunidad_observando_patrones/comercial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in commercial_loop.values[com_index][:,(0,1,3)]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


industrial_loop = gas_df[(gas_df["COMMUNITY AREA NAME"] == "Ashburn") & (gas_df["BUILDING TYPE"] == "Industrial")]
industrial_loop


for i, (census_num, (index, row)) in enumerate(zip(industrial_loop['CENSUS BLOCK'], industrial_loop[dl.gas_cols].iterrows())):
    ax0 = sns.lineplot(data=row.T, color="red")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    
    ax0.set_title(f"Census block {int(census_num)}")
    Media = row.mean()
    ax0.text(0, Media, f"{Media = :0.2e} Therms")
    
    ax0.set(ylabel="kWh")

path = images_path / Path("ensayos/clustering_gas_por_comunidad_observando_patrones/industrial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


residential_loop = gas_df[(gas_df["COMMUNITY AREA NAME"] == "Loop") & (gas_df["BUILDING TYPE"] == "Residential")]
residential_loop


f, axes = plt.subplots(5, 1, figsize=(5, 15), sharex=True)

for i, (census_num, (index, row)) in enumerate(zip(residential_loop['CENSUS BLOCK'], residential_loop[dl.gas_cols].iterrows())):
    ax0 = sns.lineplot(data=row.T, ax=axes[i], color="red")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    
    ax0.set_title(f"Census block {int(census_num)}")
    Media = row.mean()
    ax0.text(0, Media, f"{Media = :0.2e} Therms")
    
    ax0.set(ylabel="Therms")

path = images_path / Path("ensayos/clustering_gas_por_comunidad_observando_patrones/residencial")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


total_energy_by_com = dl[["COMMUNITY AREA NAME"] + dl.energy_cols + ["ELECTRICITY ACCOUNTS"]]
total_energy_by_com


total_energy_by_com[dl.energy_cols] = total_energy_by_com[dl.energy_cols].div(total_energy_by_com["ELECTRICITY ACCOUNTS"], axis=0)
del total_energy_by_com["ELECTRICITY ACCOUNTS"]
total_energy_by_com


mean_energy = total_energy_by_com.groupby('COMMUNITY AREA NAME', as_index=False).mean()
mean_energy


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = mean_energy[dl.energy_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_energia_media_en_contador/total")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(4)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,
       0, 0, 0, 3, 3, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_energia_media_en_contador/total")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = mean_energy.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    ax0.set(ylabel="kWh")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_energia_media_en_contador/total")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in mean_energy.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


zmin_score = 0.26

z_mean_energy = zscore(mean_energy[dl.energy_cols])
mean_energy_qm = mean_energy[np.logical_and(-zmin_score <= z_mean_energy, z_mean_energy <= zmin_score).all(axis=1)]
mean_energy_q1 = mean_energy[(-zmin_score > z_mean_energy).all(axis=1)]
mean_energy_q4 = mean_energy[(zmin_score < z_mean_energy).all(axis=1)]

mean_energy_qm


mean_energy_q1


mean_energy_q4


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = mean_energy_qm[dl.energy_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_qm")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(4)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([1, 3, 0, 3, 0, 3, 3, 1, 0, 0, 3, 0, 0, 3, 3, 3, 3, 0, 1, 0, 2, 0,
       0, 1, 3, 1, 1, 3, 2, 1, 0, 3, 0, 1])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_qm")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = mean_energy_qm.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    ax0.set(ylabel="kWh")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_qm")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(mean_energy_qm[dl.energy_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_qm")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in mean_energy_qm.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


kmax = 6
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = mean_energy_q1[dl.energy_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_q1")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

C:\Users\Alberto\miniconda3\envs\cdi-musiani-env\lib\site-packages\yellowbrick\utils\kneed.py:140: YellowbrickWarning:

No 'knee' or 'elbow point' detected This could be due to bad clustering, no actual clusters being formed etc.

C:\Users\Alberto\miniconda3\envs\cdi-musiani-env\lib\site-packages\yellowbrick\cluster\elbow.py:343: YellowbrickWarning:

No 'knee' or 'elbow' point detected, pass `locate_elbow=False` to remove the warning

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(3)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([1, 2, 2, 2, 2, 1, 1, 0, 2, 2, 1, 1, 2])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_q1")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = mean_energy_q1.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    ax0.set(ylabel="kWh")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    
path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_q1")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("cluster_grafica.svg"))


v = tsne(mean_energy_q1[dl.energy_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_q1")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("scatter.svg"))


v = tsne(mean_energy_q1[dl.energy_cols], 3)
fig = go.Figure()
for label in range(k):
    fig.add_trace(
        go.Scatter3d(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1],
            z=v[(labels['label']==label).values, 2]
        )
    )
fig.show()

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_q1")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("scatter3d.svg"))

<Figure size 576x396 with 0 Axes>


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in mean_energy_q1.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


kmax = 7
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = mean_energy_q4[dl.energy_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_q4")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(3)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([0, 1, 0, 0, 0, 2, 0])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_q4")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = mean_energy_q4.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} kWh")
    ax0.set(ylabel="kWh")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_q4")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("graficas.svg"))


v = tsne(mean_energy_q4[dl.energy_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_energia_media_en_contador/mean_q4")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in mean_energy_q4.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


total_gas_by_com = dl[["COMMUNITY AREA NAME"] + dl.gas_cols + ["GAS ACCOUNTS"]]
total_gas_by_com


total_gas_by_com[dl.gas_cols] = total_gas_by_com[dl.gas_cols].div(total_gas_by_com["GAS ACCOUNTS"], axis=0)
del total_gas_by_com["GAS ACCOUNTS"]
total_gas_by_com


mean_gas = total_gas_by_com.groupby('COMMUNITY AREA NAME', as_index=False).mean()
mean_gas


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = mean_gas[dl.gas_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_gas_media_en_contador/total")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(4)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 3, 3, 3, 3, 0, 3, 3, 3, 3,
       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 3, 3, 0, 3, 3, 3, 3, 1, 3, 3,
       3, 3, 3, 2, 0, 3, 3, 3, 3, 3, 3, 0, 3, 3, 3, 2, 3, 3, 3, 3, 3, 3,
       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_gas_media_en_contador/total")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = mean_gas.iloc[labels.query(f'label == {ki}').index][dl.gas_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} Therms")
    ax0.set(ylabel="Therms")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_gas_media_en_contador/total")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("graficas.svg"))


v = tsne(mean_gas[dl.gas_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_gas_media_en_contador/total")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in mean_gas.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


zmin_score = 0.26

z_mean_gas = zscore(mean_gas[dl.gas_cols])
mean_gas_qm = mean_gas[np.logical_and(-zmin_score <= z_mean_gas, z_mean_gas <= zmin_score).all(axis=1)]
mean_gas_q1 = mean_gas[(-zmin_score > z_mean_gas).all(axis=1)]
mean_gas_q4 = mean_gas[(zmin_score < z_mean_gas).all(axis=1)]

mean_gas_qm


mean_gas_q1


mean_gas_q4


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
data = mean_gas_qm[dl.gas_cols]
norm_data = scaler.fit_transform(data)

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_gas_media_en_contador/mean_qm")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(5)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

km_labels

array([4, 0, 1, 1, 3, 2, 4, 2, 1, 0, 4, 1, 1, 4])


_ = silhouette_visualizer(km_best, norm_data, colors='yellowbrick')

path = images_path / Path("ensayos/clustering_gas_media_en_contador/mean_qm")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("silhouette.svg"))

<Figure size 576x396 with 0 Axes>


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = mean_gas_qm.iloc[labels.query(f'label == {ki}').index][dl.gas_cols]
    palette={i:"darkcyan" for i in range(cluster.shape[0])}

    ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
    _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    ax0.set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    ax0.text(0, Media, f"{Media = :0.2e} Therms")
    ax0.set(ylabel="Therms")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)

path = images_path / Path("ensayos/clustering_gas_media_en_contador/mean_qm")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


v = tsne(mean_gas_qm[dl.gas_cols], 2)
for label in range(k):
    plt.scatter(
            x=v[(labels['label']==label).values, 0],
            y=v[(labels['label']==label).values, 1]
    )
plt.legend([f"Clúster {i}" for i in range(k)])

path = images_path / Path("ensayos/clustering_gas_media_en_contador/mean_qm")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("scatter.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in mean_gas_qm.values[com_index][:,0]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


f, axes = plt.subplots(6, 1, figsize=(5, 15), sharex=True)

for i, (com_name, (index, row)) in enumerate(zip(mean_gas_q1['COMMUNITY AREA NAME'], mean_gas_q1[dl.gas_cols].iterrows())):
    ax0 = sns.lineplot(data=row.T, ax=axes[i], color="red")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    
    ax0.set_title(f"Community area: {com_name}")
    Media = row.mean()
    ax0.text(0, Media, f"{Media = :0.2e} Therms")
    
    ax0.set(ylabel="Therms")

path = images_path / Path("ensayos/clustering_gas_media_en_contador/mean_q1")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


f, axes = plt.subplots(4, 1, figsize=(5, 15), sharex=True)

for i, (com_name, (index, row)) in enumerate(zip(mean_gas_q4['COMMUNITY AREA NAME'], mean_gas_q1[dl.gas_cols].iterrows())):
    ax0 = sns.lineplot(data=row.T, ax=axes[i], color="red")
    
    ax0.set_xticks([e for e in range(0, 12)])
    ax0.set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    
    ax0.set_title(f"Community area: {com_name}")
    Media = row.mean()
    ax0.text(0, Media, f"{Media = :0.2e} Therms")
    
    ax0.set(ylabel="Therms")

path = images_path / Path("ensayos/clustering_gas_media_en_contador/mean_q4")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


building_age = dl[['AVERAGE BUILDING AGE']]
building_age


f"La edad media de los edificios oscila entre {building_age.min().item()} y {building_age.max().item()} años (incluyendo filas nulas)"

'La edad media de los edificios oscila entre 0.0 y 158.0 años (incluyendo filas nulas)'


energy_building_age = dl[['AVERAGE BUILDING AGE', 'COMMUNITY AREA NAME', 'CENSUS BLOCK', 'BUILDING TYPE', 'BUILDING_SUBTYPE'] + dl.energy_cols]
gas_building_age = dl[['AVERAGE BUILDING AGE', 'COMMUNITY AREA NAME', 'CENSUS BLOCK', 'BUILDING TYPE', 'BUILDING_SUBTYPE'] + dl.gas_cols]


f"Para las filas no nulas de energía, la edad oscila entre {energy_building_age[['AVERAGE BUILDING AGE']].min().item()} y {energy_building_age[['AVERAGE BUILDING AGE']].max().item()}"

'Para las filas no nulas de energía, la edad oscila entre 0.0 y 158.0'


f"Para las filas no nulas de energía, la edad oscila entre {gas_building_age[['AVERAGE BUILDING AGE']].min().item()} y {gas_building_age[['AVERAGE BUILDING AGE']].max().item()}"

'Para las filas no nulas de energía, la edad oscila entre 0.0 y 153.5'


energy_building_age


kmax = 15
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
norm_data = scaler.fit_transform(energy_building_age[dl.energy_cols])

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_por_edad_edificio/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(5)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

np.unique(km_labels, return_counts=True)

(array([0, 1, 2, 3, 4]),
 array([65812,     2,    53,     6,   230], dtype=int64))


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = energy_building_age.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    
    palette={i:"darkcyan" for i in range(cluster.shape[0])}
    
    if len(cluster) < 10000: #para reducir el tiempo de computo no mostramos para un máximo de muestras
        ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
        _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    axes[ki].set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    axes[ki].text(0, Media, f"{Media = :0.2e} kWh")
    
    axes[ki].set_xticks([e for e in range(0, 12)])
    axes[ki].set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    axes[ki].set(ylabel="kWh")
    #print(f"Gráfica {ki} realizada")
    
    cluster_age = energy_building_age.iloc[labels.query(f'label == {ki}').index]["AVERAGE BUILDING AGE"]
    
    txt = f"Edad máxima: {cluster_age.max(): 0.2f} años\n" + \
          f"Edad mínima: {cluster_age.min(): 0.2f} años\n" + \
          f"Edad media: {cluster_age.mean(): 0.2f} años\n"
    axes[ki].text(12, Media, txt)

    
path = images_path / Path("ensayos/clustering_por_edad_edificio/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in energy_building_age.values[com_index][:,1:5]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


gas_building_age


kmax = 8
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
norm_data = scaler.fit_transform(gas_building_age[dl.gas_cols])

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_por_edad_edificio/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(4)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

np.unique(km_labels, return_counts=True)

(array([0, 1, 2, 3]), array([60314,     4,    33,   385], dtype=int64))


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = gas_building_age.iloc[labels.query(f'label == {ki}').index][dl.gas_cols]
    
    palette={i:"darkcyan" for i in range(cluster.shape[0])}
    
    if len(cluster) < 10000:
        ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
        _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    axes[ki].set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    axes[ki].text(0, Media, f"{Media = :0.2e} Therms")
    
    axes[ki].set_xticks([e for e in range(0, 12)])
    axes[ki].set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    axes[ki].set(ylabel="Therms")
    #print(f"Gráfica {ki} realizada")
    
    cluster_age = energy_building_age.iloc[labels.query(f'label == {ki}').index]["AVERAGE BUILDING AGE"]
    
    txt = f"Edad máxima: {cluster_age.max(): 0.2f} años\n" + \
          f"Edad mínima: {cluster_age.min(): 0.2f} años\n" + \
          f"Edad media: {cluster_age.mean(): 0.2f} años\n"
    axes[ki].text(12, Media, txt)

    
path = images_path / Path("ensayos/clustering_por_edad_edificio/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in gas_building_age.values[com_index][:,1:5]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in comunities_by_cluster[k]:
        print("\t", e)


housesize = dl[['AVERAGE HOUSESIZE', 'BUILDING TYPE']]
housesize = housesize[housesize['BUILDING TYPE'] == "Residential"]
housesize


minsize = housesize['AVERAGE HOUSESIZE'].min().item()
maxsize = housesize['AVERAGE HOUSESIZE'].max().item()

f'El valor de dimensión del hogar oscila entre {minsize} y {maxsize}'

'El valor de dimensión del hogar oscila entre 0.0 y 9.0'


energy_housesize = dl[dl.energy_cols + ['AVERAGE HOUSESIZE', 'BUILDING TYPE', 'COMMUNITY AREA NAME']]
energy_housesize = energy_housesize[energy_housesize['BUILDING TYPE'] == "Residential"][dl.energy_cols + ['AVERAGE HOUSESIZE', 'COMMUNITY AREA NAME']]
energy_housesize


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
norm_data = scaler.fit_transform(energy_housesize[dl.energy_cols])

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_por_dimension_hogar/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(5)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

np.unique(km_labels, return_counts=True)

(array([0, 1, 2, 3, 4]),
 array([17752,     4,  1473,    46, 30165], dtype=int64))


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = energy_housesize.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    
    palette={i:"darkcyan" for i in range(cluster.shape[0])}
    
    if len(cluster) < 10000:
        ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
        _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    axes[ki].set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    axes[ki].text(0, Media, f"{Media = :0.2e} kWh")
    
    axes[ki].set_xticks([e for e in range(0, 12)])
    axes[ki].set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    axes[ki].set(ylabel="kWh")
    #print(f"Gráfica {ki} realizada")
    
    cluster_housesize = energy_housesize.iloc[labels.query(f'label == {ki}').index]["AVERAGE HOUSESIZE"]
    
    txt = f"Max housesize: {cluster_housesize.max()}\n" + \
          f"Min housesize: {cluster_housesize.min()}\n" + \
          f"Mean housesize:{cluster_housesize.mean()}"
    
    axes[ki].text(12, Media, txt)

path = images_path / Path("ensayos/clustering_por_dimension_hogar/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in energy_housesize.values[com_index][:,-1]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in np.unique(comunities_by_cluster[k]):
        print("\t", e)


gas_housesize = dl[dl.gas_cols + ['AVERAGE HOUSESIZE', 'BUILDING TYPE', 'COMMUNITY AREA NAME']]
gas_housesize = gas_housesize[gas_housesize['BUILDING TYPE'] == "Residential"][dl.gas_cols + ['AVERAGE HOUSESIZE', 'COMMUNITY AREA NAME']]
gas_housesize


kmax = 10
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
norm_data = scaler.fit_transform(gas_housesize[dl.gas_cols])

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_por_dimension_hogar/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(4)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

np.unique(km_labels, return_counts=True)

(array([0, 1, 2, 3]), array([17451, 26212,    68,  2464], dtype=int64))


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = gas_housesize.iloc[labels.query(f'label == {ki}').index][dl.gas_cols]
    
    palette={i:"darkcyan" for i in range(cluster.shape[0])}
    
    if len(cluster) < 10000:
        ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
        _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    axes[ki].set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    axes[ki].text(0, Media, f"{Media = :0.2e} Therms")
    
    axes[ki].set_xticks([e for e in range(0, 12)])
    axes[ki].set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    axes[ki].set(ylabel="Therms")
    #print(f"Gráfica {ki} realizada")
    
    cluster_housesize = gas_housesize.iloc[labels.query(f'label == {ki}').index]["AVERAGE HOUSESIZE"]
    
    txt = f"Max housesize: {cluster_housesize.max()}\n" + \
          f"Min housesize: {cluster_housesize.min()}\n" + \
          f"Mean housesize:{cluster_housesize.mean()}"
    
    axes[ki].text(12, Media, txt)

path = images_path / Path("ensayos/clustering_por_dimension_hogar/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in gas_housesize.values[com_index][:, -1]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in np.unique(comunities_by_cluster[k]):
        print("\t", e)


energy_occupied = dl[dl.energy_cols + ['OCCUPIED UNITS PERCENTAGE', 'BUILDING TYPE', 'COMMUNITY AREA NAME']]
energy_occupied = energy_occupied[energy_occupied['BUILDING TYPE'] == "Residential"][dl.energy_cols + ['OCCUPIED UNITS PERCENTAGE', 'COMMUNITY AREA NAME']]
energy_occupied


kmax = 15
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
norm_data = scaler.fit_transform(energy_occupied[dl.energy_cols])

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_por_ocupacion_total/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(5)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

np.unique(km_labels, return_counts=True)

(array([0, 1, 2, 3, 4]),
 array([29584, 17577,     3,    46,  1451], dtype=int64))


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = energy_occupied.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    
    palette={i:"darkcyan" for i in range(cluster.shape[0])}
    
    if len(cluster) < 10000:
        ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
        _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    axes[ki].set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    axes[ki].text(0, Media, f"{Media = :0.2e} kWh")
    
    axes[ki].set_xticks([e for e in range(0, 12)])
    axes[ki].set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    axes[ki].set(ylabel="kWh")
    #print(f"Gráfica {ki} realizada")
    
    cluster_occupied = energy_occupied.iloc[labels.query(f'label == {ki}').index]["OCCUPIED UNITS PERCENTAGE"]
    
    txt = f"Porcentaje ocupación máxima: {cluster_occupied.max()}\n" + \
          f"Porcentaje ocupación minimo: {cluster_occupied.min()}\n" + \
          f"Porcentaje ocupación medio:{cluster_occupied.mean()}"
    
    axes[ki].text(12, Media, txt)

    
path = images_path / Path("ensayos/clustering_por_ocupacion_total/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in energy_occupied.values[com_index][:, -1]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in np.unique(comunities_by_cluster[k]):
        print("\t", e)


gas_occupied = dl[dl.gas_cols + ['OCCUPIED UNITS PERCENTAGE', 'BUILDING TYPE', 'COMMUNITY AREA NAME']]
gas_occupied = gas_occupied[gas_occupied['BUILDING TYPE'] == "Residential"][dl.gas_cols + ['OCCUPIED UNITS PERCENTAGE', 'COMMUNITY AREA NAME']]
gas_occupied


kmax = 15
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
norm_data = scaler.fit_transform(gas_occupied[dl.gas_cols])

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_por_ocupacion_total/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(5)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

np.unique(km_labels, return_counts=True)

(array([0, 1, 2, 3, 4]),
 array([23267,  4440,    29,   483, 17295], dtype=int64))


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = gas_occupied.iloc[labels.query(f'label == {ki}').index][dl.gas_cols]
    
    palette={i:"darkcyan" for i in range(cluster.shape[0])}
    
    if len(cluster) < 1000:
        ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
        _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    axes[ki].set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    axes[ki].text(0, Media, f"{Media = :0.2e} Therms")
    
    axes[ki].set_xticks([e for e in range(0, 12)])
    axes[ki].set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    axes[ki].set(ylabel="Therms")
    #print(f"Gráfica {ki} realizada")
    
    cluster_occupied = gas_occupied.iloc[labels.query(f'label == {ki}').index]["OCCUPIED UNITS PERCENTAGE"]
    
    txt = f"Porcentaje ocupación máxima: {cluster_occupied.max()}\n" + \
          f"Porcentaje ocupación minimo: {cluster_occupied.min()}\n" + \
          f"Porcentaje ocupación medio:{cluster_occupied.mean()}"
    
    axes[ki].text(12, Media, txt)

path = images_path / Path("ensayos/clustering_por_ocupacion_total/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in gas_occupied.values[com_index][:, -1]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in np.unique(comunities_by_cluster[k]):
        print("\t", e)


energy_occupied = dl[dl.energy_cols + ['RENTER-OCCUPIED HOUSING PERCENTAGE', 'BUILDING TYPE', 'COMMUNITY AREA NAME']]
energy_occupied = energy_occupied[energy_occupied['BUILDING TYPE'] == "Residential"][dl.energy_cols + ['RENTER-OCCUPIED HOUSING PERCENTAGE', 'COMMUNITY AREA NAME']]
energy_occupied


kmax = 15
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
norm_data = scaler.fit_transform(energy_occupied[dl.energy_cols])

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_por_ocupacion_en_renta/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(5)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

np.unique(km_labels, return_counts=True)

(array([0, 1, 2, 3, 4]),
 array([29677, 17501,     3,    46,  1374], dtype=int64))


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = energy_occupied.iloc[labels.query(f'label == {ki}').index][dl.energy_cols]
    
    palette={i:"darkcyan" for i in range(cluster.shape[0])}
    
    if len(cluster) < 10000:
        ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
        _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    axes[ki].set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    axes[ki].text(0, Media, f"{Media = :0.2e} kWh")
    
    axes[ki].set_xticks([e for e in range(0, 12)])
    axes[ki].set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    axes[ki].set(ylabel="kWh")
    #print(f"Gráfica {ki} realizada")
    
    cluster_occupied = energy_occupied.iloc[labels.query(f'label == {ki}').index]["RENTER-OCCUPIED HOUSING PERCENTAGE"]
    
    txt = f"Porcentaje ocupación de casas en renta máxima: {cluster_occupied.max()}\n" + \
          f"Porcentaje ocupación de casas en renta minimo: {cluster_occupied.min()}\n" + \
          f"Porcentaje ocupación de casas en renta medio:{cluster_occupied.mean()}"
    
    axes[ki].text(12, Media, txt)

path = images_path / Path("ensayos/clustering_por_ocupacion_en_renta/energia")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in energy_occupied.values[com_index][:, -1]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in np.unique(comunities_by_cluster[k]):
        print("\t", e)


gas_occupied = dl[dl.gas_cols + ['RENTER-OCCUPIED HOUSING PERCENTAGE', 'BUILDING TYPE', 'COMMUNITY AREA NAME']]
gas_occupied = gas_occupied[gas_occupied['BUILDING TYPE'] == "Residential"][dl.gas_cols + ['RENTER-OCCUPIED HOUSING PERCENTAGE', 'COMMUNITY AREA NAME']]
gas_occupied


kmax = 15
k_values = (2, kmax)

scaler = preprocessing.StandardScaler()
norm_data = scaler.fit_transform(gas_occupied[dl.gas_cols])

km = KMeansCluster(norm_data)
km.cluster(k_values)

path = images_path / Path("ensayos/clustering_por_ocupacion_en_renta/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("elbow.svg"))

<Figure size 576x396 with 0 Axes>


km_best = km.get_kmeans_of(5)
km_labels = km_best.labels_
km_centroids = scaler.inverse_transform(km_best.cluster_centers_)

np.unique(km_labels, return_counts=True)

(array([0, 1, 2, 3, 4]),
 array([  200, 24350,  3402, 17495,    16], dtype=int64))


labels = pd.DataFrame(km_labels, columns=["label"])

k = km_centroids.shape[0]

f, axes = plt.subplots(k, 1, figsize=(5, 15), sharex=True)

for ki in range(k):
    cluster = gas_occupied.iloc[labels.query(f'label == {ki}').index][dl.gas_cols]
    
    palette={i:"darkcyan" for i in range(cluster.shape[0])}
    
    if len(cluster) < 1000:
        ax0 = sns.lineplot(data=cluster.values.T, ax=axes[ki], palette=palette, alpha=0.3)
        _ = [line.set_linestyle("-") for line in ax0.lines]
    
    ax0c = sns.lineplot(data=km_centroids[ki, :].T, ax=axes[ki], color="red")
    _ = [line.set_linestyle("-") for line in ax0c.lines]
    
    ax0c.legend(ax0c.lines[-1:], ["Centroide"]);
    
    axes[ki].set_title(f"Clúster {ki}")
    Media = km_centroids[ki, :].mean()
    axes[ki].text(0, Media, f"{Media = :0.2e} Therms")
    
    axes[ki].set_xticks([e for e in range(0, 12)])
    axes[ki].set_xticklabels([f"{datetime.date(2008, i, 1).strftime('%B').upper()}" for i in range(1, 13)], rotation=90)
    axes[ki].set(ylabel="Therms")
    #print(f"Gráfica {ki} realizada")
    
    cluster_occupied = gas_occupied.iloc[labels.query(f'label == {ki}').index]["RENTER-OCCUPIED HOUSING PERCENTAGE"]
    
    txt = f"Porcentaje ocupación de casas en renta máxima: {cluster_occupied.max()}\n" + \
          f"Porcentaje ocupación de casas en renta minimo: {cluster_occupied.min()}\n" + \
          f"Porcentaje ocupación de casas en renta medio:{cluster_occupied.mean()}"
    
    axes[ki].text(12, Media, txt)

path = images_path / Path("ensayos/clustering_por_ocupacion_en_renta/gas")
path.mkdir(parents=True, exist_ok=True)
plt.savefig(path / Path("grafica.svg"))


comunities_by_cluster = {}

for nlab in range(k):
    com_index = np.where(labels==nlab)[0]
    com_list = []
    for c in gas_occupied.values[com_index][:, -1]:
        com_list.append(c)
    comunities_by_cluster[nlab] = com_list

for k in sorted(comunities_by_cluster, key=lambda k: len(comunities_by_cluster[k])):
    print(f"Al clúster {k} pertenecen las comunidades:")
    for e in np.unique(comunities_by_cluster[k]):
        print("\t", e)

	COMMUNITY AREA NAME	CENSUS BLOCK	BUILDING TYPE	BUILDING_SUBTYPE	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	...	TOTAL POPULATION	TOTAL UNITS	AVERAGE STORIES	AVERAGE BUILDING AGE	AVERAGE HOUSESIZE	OCCUPIED UNITS	OCCUPIED UNITS PERCENTAGE	RENTER-OCCUPIED HOUSING UNITS	RENTER-OCCUPIED HOUSING PERCENTAGE	OCCUPIED HOUSING UNITS
0	Archer Heights	1.703157e+14	Residential	Multi < 7	NaN	NaN	NaN	NaN	NaN	NaN	...	89.0	24.0	2.00	71.33	3.87	23.0	0.9582	9.0	0.3910	23.0
1	Ashburn	1.703170e+14	Residential	Multi 7+	7334.0	7741.0	4214.0	4284.0	2518.0	4273.0	...	112.0	67.0	2.00	41.00	1.81	62.0	0.9254	50.0	0.8059	62.0
2	Auburn Gresham	1.703171e+14	Commercial	Multi < 7	NaN	NaN	NaN	NaN	NaN	NaN	...	102.0	48.0	3.00	86.00	3.00	34.0	0.7082	23.0	0.6759	34.0
3	Austin	1.703125e+14	Commercial	Multi < 7	NaN	NaN	NaN	NaN	NaN	NaN	...	121.0	56.0	2.00	84.00	2.95	41.0	0.7321	32.0	0.7800	41.0
4	Austin	1.703125e+14	Commercial	Multi < 7	NaN	NaN	NaN	NaN	NaN	NaN	...	62.0	23.0	2.00	85.00	3.26	19.0	0.8261	11.0	0.5790	19.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
67046	Woodlawn	1.703184e+14	Residential	Single Family	2705.0	1318.0	1582.0	1465.0	1494.0	2990.0	...	116.0	55.0	1.00	0.00	3.14	37.0	0.6727	26.0	0.7030	37.0
67047	Woodlawn	1.703184e+14	Commercial	Multi < 7	1005.0	1760.0	1521.0	1832.0	2272.0	2361.0	...	31.0	24.0	3.00	104.50	2.07	15.0	0.6250	13.0	0.8670	15.0
67048	Woodlawn	1.703184e+14	Residential	Multi < 7	3567.0	3031.0	2582.0	2295.0	7902.0	4987.0	...	31.0	24.0	2.33	100.67	2.07	15.0	0.6250	13.0	0.8670	15.0
67049	Woodlawn	1.703184e+14	Residential	Single Family	1208.0	1055.0	1008.0	1109.0	1591.0	1367.0	...	0.0	0.0	1.00	0.00	0.00	0.0	NaN	0.0	NaN	0.0
67050	Woodlawn	1.703184e+14	Residential	Multi < 7	2717.0	3057.0	2695.0	3793.0	4237.0	5383.0	...	77.0	49.0	2.00	79.40	2.57	30.0	0.6122	28.0	0.9329	30.0

	COMMUNITY AREA NAME	BUILDING TYPE	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
0	Albany Park	Commercial	6485.973190	6598.123324	6350.994638	6565.369973	8521.268097	10648.798928	10824.525469	9204.222520	7193.860590	7063.611260	8875.603217	9536.718499
1	Albany Park	Industrial	4387.000000	3671.000000	4849.000000	4923.000000	6580.000000	8196.000000	7074.000000	7556.000000	4655.000000	6193.000000	12597.000000	8237.000000
2	Albany Park	Residential	4493.482445	4868.560468	4831.300390	4704.583875	6202.847854	9507.514954	10494.762029	7880.837451	5533.161248	5713.767230	8412.635891	9043.319896
3	Archer Heights	Commercial	76154.428571	77902.155844	74525.714286	72982.935065	76074.220779	85996.129870	84604.389610	85607.376623	77916.636364	73602.805195	76310.571429	84052.545455
4	Archer Heights	Industrial	1100.000000	634.000000	508.000000	456.000000	552.000000	829.000000	4405.000000	1304.000000	895.000000	657.000000	784.000000	1586.000000
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
168	West Town	Commercial	7890.108716	8219.340206	7867.381443	7907.283037	9768.865979	12332.753515	13705.432052	13291.417057	10410.490159	9151.462980	10051.944705	11954.955014
169	West Town	Industrial	7642.000000	6762.000000	5174.000000	6681.000000	8466.000000	16203.000000	16007.000000	13908.000000	9200.000000	7256.000000	8621.000000	10312.000000
170	West Town	Residential	4035.878754	4627.875921	4392.944476	4392.706516	5338.343909	7254.643626	9278.979603	9354.343909	7011.481020	5714.180737	6840.808499	8876.675921
171	Woodlawn	Commercial	66321.169231	62376.574359	46025.943590	44388.046154	41310.712821	48550.261538	44273.635897	52702.687179	59037.825641	50860.584615	53246.569231	76645.671795
172	Woodlawn	Residential	4064.865482	4168.939086	4120.944162	4000.304569	4449.756345	5899.697970	7016.401015	6889.312183	5007.302030	5081.281726	6718.832487	7993.091371

	COMMUNITY AREA NAME	BUILDING TYPE	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
0	Albany Park	Commercial	6485.973190	6598.123324	6350.994638	6565.369973	8521.268097	10648.798928	10824.525469	9204.222520	7193.860590	7063.611260	8875.603217	9536.718499
3	Archer Heights	Commercial	76154.428571	77902.155844	74525.714286	72982.935065	76074.220779	85996.129870	84604.389610	85607.376623	77916.636364	73602.805195	76310.571429	84052.545455
6	Armour Square	Commercial	37698.041667	34363.145833	30411.239583	31034.666667	31278.093750	33461.270833	42105.614583	39720.145833	32412.708333	23017.114583	31733.833333	48211.739583
9	Ashburn	Commercial	24129.737374	22406.404040	20967.484848	21031.595960	26618.222222	28948.676768	29027.707071	26762.808081	22696.373737	21852.020202	25317.878788	28463.444444
11	Auburn Gresham	Commercial	5184.100386	5056.308880	5084.494208	5122.316602	6511.660232	7688.424710	7925.613900	6999.822394	5477.868726	5558.455598	6506.455598	7659.637066
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
162	West Lawn	Commercial	65512.178947	71017.073684	67999.989474	69908.431579	84869.778947	94092.010526	91777.000000	92741.736842	78164.905263	66165.336842	69139.621053	78315.905263
164	West Pullman	Commercial	4018.162791	4657.023256	4679.779070	4860.197674	5639.651163	6886.476744	7874.616279	7650.453488	6234.651163	5450.593023	6107.093023	7010.883721
166	West Ridge	Commercial	9307.416335	9402.619522	8822.699203	9619.591633	12595.511952	15000.701195	14534.021912	12644.521912	9959.976096	10351.906375	12838.211155	13003.069721
168	West Town	Commercial	7890.108716	8219.340206	7867.381443	7907.283037	9768.865979	12332.753515	13705.432052	13291.417057	10410.490159	9151.462980	10051.944705	11954.955014
171	Woodlawn	Commercial	66321.169231	62376.574359	46025.943590	44388.046154	41310.712821	48550.261538	44273.635897	52702.687179	59037.825641	50860.584615	53246.569231	76645.671795

	COMMUNITY AREA NAME	BUILDING TYPE	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
2	Albany Park	Residential	4493.482445	4868.560468	4831.300390	4704.583875	6202.847854	9507.514954	10494.762029	7880.837451	5533.161248	5713.767230	8412.635891	9043.319896
5	Archer Heights	Residential	4779.779605	3830.042763	3381.667763	4259.944079	5050.868421	7143.690789	10079.098684	8852.562500	5186.148026	4123.243421	5542.542763	7024.115132
8	Armour Square	Residential	4880.892655	5422.807910	5269.322034	5195.807910	5739.259887	6977.977401	10876.067797	10688.847458	8135.774011	6055.570621	7107.508475	9459.282486
10	Ashburn	Residential	7685.557441	7535.497389	7290.016971	7476.365535	10381.006527	16359.704961	18197.938642	13426.244125	9015.569191	8341.000000	12181.706266	13469.849869
12	Auburn Gresham	Residential	5250.104076	5428.169991	5243.229835	5094.189939	6166.013877	8637.880312	9349.874241	7706.798786	5858.764094	6152.209887	8788.036427	10000.882914
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
163	West Lawn	Residential	5884.403716	5566.481419	5527.908784	5224.923986	7577.138514	11927.521959	13929.736486	11061.168919	6631.606419	5928.351351	7740.572635	9650.545608
165	West Pullman	Residential	2512.937376	4048.336978	3890.841948	4105.110338	4580.604374	5652.351889	7412.451292	7063.981113	5111.528827	4619.525845	5552.450298	7356.383698
167	West Ridge	Residential	5471.784787	5573.705937	5368.393321	5863.319109	7990.311688	11648.390538	11575.739332	8624.476809	5947.495362	6420.602041	9243.713358	9150.877551
170	West Town	Residential	4035.878754	4627.875921	4392.944476	4392.706516	5338.343909	7254.643626	9278.979603	9354.343909	7011.481020	5714.180737	6840.808499	8876.675921
172	Woodlawn	Residential	4064.865482	4168.939086	4120.944162	4000.304569	4449.756345	5899.697970	7016.401015	6889.312183	5007.302030	5081.281726	6718.832487	7993.091371

	COMMUNITY AREA NAME	BUILDING TYPE	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
1	Albany Park	Industrial	4.387000e+03	3.671000e+03	4849.0	4.923000e+03	6.580000e+03	8.196000e+03	7.074000e+03	7.556000e+03	4.655000e+03	6.193000e+03	1.259700e+04	8.237000e+03
4	Archer Heights	Industrial	1.100000e+03	6.340000e+02	508.0	4.560000e+02	5.520000e+02	8.290000e+02	4.405000e+03	1.304000e+03	8.950000e+02	6.570000e+02	7.840000e+02	1.586000e+03
7	Armour Square	Industrial	2.537600e+04	2.299100e+04	18723.0	2.217700e+04	3.171000e+04	3.259000e+04	6.834600e+04	5.245100e+04	3.563400e+04	2.720200e+04	2.700300e+04	3.384400e+04
14	Austin	Industrial	4.581100e+04	4.222500e+04	34700.0	4.234500e+04	3.812000e+04	4.082000e+04	4.624200e+04	5.580700e+04	4.414400e+04	4.315900e+04	4.946300e+04	5.294800e+04
19	Avondale	Industrial	1.364200e+04	1.302200e+04	11056.0	1.025400e+04	1.116000e+04	1.404700e+04	1.264000e+04	1.122500e+04	1.078100e+04	1.049000e+04	1.267200e+04	1.496800e+04
28	Brighton Park	Industrial	0.000000e+00	0.000000e+00	0.0	0.000000e+00	0.000000e+00	5.440000e+02	9.760000e+02	7.620000e+02	1.137000e+03	4.780000e+02	6.360000e+02	7.020000e+02
49	Edgewater	Industrial	3.373000e+03	2.297500e+03	2023.5	3.184500e+03	4.155000e+03	5.877000e+03	5.980000e+03	4.898000e+03	3.381500e+03	2.712000e+03	4.574000e+03	3.878000e+03
56	Forest Glen	Industrial	4.909000e+03	4.079000e+03	6355.0	5.398000e+03	9.727000e+03	1.626400e+04	1.414400e+04	8.681000e+03	5.932000e+03	4.960000e+03	7.843000e+03	6.714000e+03
61	Gage Park	Industrial	4.140000e+02	8.370000e+02	1083.0	1.182000e+03	8.420000e+02	1.140000e+03	1.468000e+03	3.129000e+03	1.752000e+03	1.122000e+03	1.156000e+03	1.213000e+03
72	Hermosa	Industrial	7.989000e+03	1.101300e+04	10109.0	9.970000e+03	9.236000e+03	1.098200e+04	1.255600e+04	1.058800e+04	9.284000e+03	8.591000e+03	1.145800e+04	3.539000e+04
79	Irving Park	Industrial	1.262500e+03	1.201000e+03	1081.0	9.265000e+02	1.410000e+03	2.742500e+03	2.670000e+03	2.321500e+03	1.177000e+03	1.290500e+03	1.844000e+03	2.265500e+03
84	Kenwood	Industrial	5.748000e+03	5.309000e+03	5134.0	5.234000e+03	6.408000e+03	8.681000e+03	1.065400e+04	1.018900e+04	7.336000e+03	5.350000e+03	5.629000e+03	5.372000e+03
87	Lakeview	Industrial	4.662500e+03	4.486500e+03	4072.5	6.243000e+03	7.026000e+03	8.059500e+03	7.757000e+03	8.133000e+03	6.719500e+03	5.346000e+03	5.262500e+03	6.658500e+03
94	Logan Square	Industrial	1.196000e+03	1.311333e+03	1283.0	1.115667e+03	1.338333e+03	1.713000e+03	2.599667e+03	2.688000e+03	1.659667e+03	1.460667e+03	1.656333e+03	2.131000e+03
111	Near South Side	Industrial	4.288200e+04	1.323340e+05	126654.0	1.296920e+05	8.660800e+04	9.994500e+04	1.491320e+05	1.419530e+05	1.091400e+05	8.356000e+04	1.187180e+05	1.740390e+05
114	Near West Side	Industrial	7.081809e+06	7.030604e+06	5381515.0	5.780886e+06	5.327624e+06	6.754392e+06	5.782397e+06	6.184596e+06	6.439283e+06	5.710067e+06	6.235068e+06	8.365907e+06
121	North Lawndale	Industrial	0.000000e+00	7.670000e+03	3828.0	1.701000e+03	3.416000e+03	4.208000e+03	4.714000e+03	5.175000e+03	4.363000e+03	3.994000e+03	6.656000e+03	1.142300e+04
146	South Lawndale	Industrial	2.742770e+05	2.763710e+05	246830.0	2.764630e+05	2.286560e+05	3.336580e+05	3.289420e+05	3.336430e+05	2.802420e+05	3.353490e+05	2.643730e+05	2.832740e+05
169	West Town	Industrial	7.642000e+03	6.762000e+03	5174.0	6.681000e+03	8.466000e+03	1.620300e+04	1.600700e+04	1.390800e+04	9.200000e+03	7.256000e+03	8.621000e+03	1.031200e+04

	COMMUNITY AREA NAME	BUILDING TYPE	THERM JANUARY 2010	THERM FEBRUARY 2010	THERM MARCH 2010	THERM APRIL 2010	THERM MAY 2010	THERM JUNE 2010	THERM JULY 2010	THERM AUGUST 2010	THERM SEPTEMBER 2010	THERM OCTOBER 2010	THERM NOVEMBER 2010	THERM DECEMBER 2010
0	Albany Park	Commercial	2452.952909	2075.800554	1740.199446	959.753463	651.722992	428.936288	383.991690	338.675900	371.581717	427.340720	735.626039	1650.426593
1	Albany Park	Industrial	2146.000000	1963.000000	1967.000000	1038.000000	681.000000	475.000000	336.000000	344.000000	486.000000	549.000000	621.000000	1227.000000
2	Albany Park	Residential	3241.642670	2769.778796	2315.806283	1258.925393	809.119110	453.017016	340.260471	299.039267	310.349476	378.269634	804.899215	2178.858639
3	Archer Heights	Commercial	8188.828571	6049.671429	4799.785714	2470.214286	1610.285714	1260.871429	1232.642857	1075.342857	1031.357143	1411.228571	2701.742857	6576.000000
4	Archer Heights	Industrial	955.000000	815.000000	713.000000	586.000000	442.000000	367.000000	363.000000	328.000000	330.000000	330.000000	402.000000	651.000000
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
169	West Town	Commercial	1986.941355	1696.244692	1459.205258	727.702730	455.524772	327.140546	277.324570	253.211325	268.819009	318.184024	540.920121	1413.849343
170	West Town	Industrial	1579.000000	1517.000000	944.000000	374.000000	299.000000	227.000000	224.000000	211.000000	219.000000	254.000000	491.000000	1720.000000
171	West Town	Residential	2513.479660	2132.196114	1986.321190	969.794171	573.477231	359.237401	251.375228	227.674560	234.496053	285.306618	536.063752	1626.820886
172	Woodlawn	Commercial	2834.130682	2401.289773	2113.295455	1213.556818	792.159091	471.852273	310.397727	258.914773	335.318182	500.318182	994.585227	1930.295455
173	Woodlawn	Residential	2977.754569	2526.342037	2466.485640	1323.885117	802.026110	462.357702	273.563969	218.070496	250.334204	345.509138	767.381201	1826.772846

	COMMUNITY AREA NAME	BUILDING TYPE	THERM JANUARY 2010	THERM FEBRUARY 2010	THERM MARCH 2010	THERM APRIL 2010	THERM MAY 2010	THERM JUNE 2010	THERM JULY 2010	THERM AUGUST 2010	THERM SEPTEMBER 2010	THERM OCTOBER 2010	THERM NOVEMBER 2010	THERM DECEMBER 2010
1	Albany Park	Industrial	2146.0	1963.000000	1967.000000	1038.000000	681.000000	475.000000	336.0	344.000000	486.000000	549.000000	621.000000	1227.000000
4	Archer Heights	Industrial	955.0	815.000000	713.000000	586.000000	442.000000	367.000000	363.0	328.000000	330.000000	330.000000	402.000000	651.000000
9	Ashburn	Industrial	930.0	940.000000	978.000000	368.000000	246.000000	104.000000	11.0	18.000000	11.000000	18.000000	126.000000	380.000000
14	Austin	Industrial	18043.5	17736.000000	15444.500000	6863.000000	5171.000000	2614.500000	3016.0	2598.000000	3047.000000	4616.000000	7567.500000	14517.000000
27	Brighton Park	Industrial	331.0	66.000000	205.000000	93.000000	65.000000	43.000000	28.0	33.000000	19.000000	38.000000	86.000000	191.000000
46	East Side	Industrial	2587.0	2152.000000	1263.000000	404.000000	242.000000	7.000000	6.0	6.000000	6.000000	40.000000	260.000000	1759.000000
49	Edgewater	Industrial	797.0	749.000000	511.500000	232.000000	148.500000	93.500000	78.0	84.000000	102.000000	117.500000	272.000000	844.500000
56	Forest Glen	Industrial	1126.0	921.000000	592.000000	362.000000	228.000000	160.000000	243.0	192.000000	179.000000	230.000000	396.000000	1046.000000
61	Gage Park	Industrial	24.0	30.000000	25.000000	15.000000	24.000000	15.000000	19.0	12.000000	13.000000	14.000000	12.000000	18.000000
72	Hermosa	Industrial	2157.0	1434.000000	1022.000000	499.000000	152.000000	55.000000	46.0	28.000000	25.000000	102.000000	248.000000	1544.000000
79	Irving Park	Industrial	659.0	579.500000	367.500000	143.500000	62.000000	30.000000	19.0	16.000000	15.500000	29.500000	86.500000	492.500000
82	Jefferson Park	Industrial	3328.0	3034.000000	2073.000000	1140.500000	317.500000	75.000000	39.0	38.500000	37.000000	124.500000	775.500000	2094.000000
85	Kenwood	Industrial	1179.0	2306.000000	1878.000000	957.000000	768.000000	776.000000	660.0	729.000000	811.000000	767.000000	813.000000	2251.000000
88	Lakeview	Industrial	13290.0	11153.500000	10964.500000	8081.500000	7949.500000	7089.500000	8063.0	7671.000000	7952.500000	8158.500000	8921.000000	13633.000000
111	Near South Side	Industrial	13111.0	10778.000000	8246.000000	3034.000000	12429.000000	249.000000	288.0	252.000000	276.000000	3281.000000	4919.000000	10904.000000
114	Near West Side	Industrial	9963.0	8585.857143	5366.142857	2600.714286	1314.571429	1281.857143	218.0	269.714286	289.428571	1161.428571	3826.428571	7932.285714
121	North Lawndale	Industrial	2505.0	3313.000000	2372.000000	787.000000	150.000000	38.000000	86.0	38.000000	8.000000	156.000000	100.000000	1786.000000
126	Norwood Park	Industrial	1120.0	888.000000	797.000000	406.000000	191.000000	28.000000	2.0	1.000000	5.000000	46.000000	269.000000	690.000000
147	South Lawndale	Industrial	8450.0	7489.000000	7957.000000	4860.000000	1718.000000	17.000000	17.0	637.000000	18.000000	195.000000	2767.000000	4067.000000
170	West Town	Industrial	1579.0	1517.000000	944.000000	374.000000	299.000000	227.000000	224.0	211.000000	219.000000	254.000000	491.000000	1720.000000

	COMMUNITY AREA NAME	CENSUS BLOCK	BUILDING TYPE	BUILDING_SUBTYPE	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
120	Loop	1.703184e+14	Commercial	Multi 7+	73300.0	73177.0	55407.0	40004.0	41641.0	46163.0	58720.0	52091.0	53757.0	48746.0	75111.0	119320.0
121	Loop	1.703184e+14	Commercial	Commercial	1937016.0	1612999.0	1378571.0	1299736.0	1376510.0	1453082.0	1448453.0	1373392.0	1313078.0	1180224.0	2100492.0	2229423.0
354	Loop	1.703132e+14	Commercial	Commercial	1931433.0	1702836.0	1708126.0	1922289.0	2076298.0	1972560.0	2338763.0	2375021.0	2122579.0	1554759.0	1863090.0	2003699.0
945	Loop	1.703184e+14	Commercial	Commercial	2370788.0	2119876.0	2082421.0	2087045.0	2270125.0	3095521.0	3011561.0	2884957.0	2371234.0	2128850.0	2341563.0	2729045.0
946	Loop	1.703184e+14	Commercial	Municipal	1775651.0	1433137.0	1393901.0	1171420.0	1397811.0	1562200.0	1565775.0	1400773.0	1353486.0	1327522.0	1697576.0	2130009.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
38591	Loop	1.703184e+14	Commercial	Commercial	306959.0	280224.0	134944.0	95925.0	132642.0	150710.0	219940.0	187643.0	173423.0	137245.0	239216.0	446574.0
38592	Loop	1.703184e+14	Commercial	Commercial	42600.0	485192.0	510282.0	522265.0	607110.0	510410.0	593315.0	620349.0	557672.0	512367.0	514068.0	463212.0
38593	Loop	1.703184e+14	Commercial	Commercial	247181.0	209955.0	213677.0	220961.0	303981.0	308345.0	388278.0	370715.0	294093.0	242598.0	203601.0	237975.0
38594	Loop	1.703184e+14	Commercial	Multi 7+	50822.0	46418.0	40971.0	35079.0	47196.0	53944.0	77830.0	66578.0	62182.0	41404.0	46031.0	61226.0
38595	Loop	1.703184e+14	Commercial	Commercial	51003.0	34568.0	29774.0	24527.0	32431.0	37194.0	51768.0	42745.0	35956.0	26728.0	30434.0	39684.0

	COMMUNITY AREA NAME	CENSUS BLOCK	BUILDING TYPE	BUILDING_SUBTYPE	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
21904	Loop	1.703132e+14	Residential	Multi < 7	4576.0	5264.0	4321.0	3977.0	5025.0	7709.0	8728.0	8350.0	9765.0	6754.0	6048.0	6631.0
38489	Loop	1.703132e+14	Residential	Multi 7+	12625.0	12072.0	11304.0	10223.0	12654.0	13221.0	19326.0	15676.0	11690.0	9369.0	13157.0	14402.0
38491	Loop	1.703132e+14	Residential	Multi 7+	31735.0	31549.0	30020.0	27812.0	38957.0	43405.0	57833.0	54271.0	39998.0	35803.0	34729.0	44797.0
38501	Loop	1.703132e+14	Residential	Multi 7+	150389.0	156138.0	153203.0	162340.0	223682.0	240178.0	329181.0	309317.0	224617.0	194145.0	221436.0	269572.0
38513	Loop	1.703184e+14	Residential	Single Family	690.0	887.0	1248.0	881.0	1724.0	1910.0	2020.0	1439.0	620.0	371.0	2386.0	1112.0

	COMMUNITY AREA NAME	CENSUS BLOCK	BUILDING TYPE	BUILDING_SUBTYPE	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
16299	Near West Side	1.703128e+14	Industrial	Industrial	21214017.0	21065500.0	16121105.0	17310058.0	15938553.0	20209197.0	17287870.0	18493853.0	19280342.0	17100282.0	18671279.0	25060008.0
42543	Near West Side	1.703183e+14	Industrial	Industrial	25445.0	22534.0	20547.0	30382.0	40722.0	48960.0	51211.0	53044.0	34103.0	27225.0	29469.0	32424.0
42571	Near West Side	1.703183e+14	Industrial	Industrial	5966.0	3779.0	2893.0	2219.0	3597.0	5018.0	8110.0	6890.0	3403.0	2693.0	4456.0	5290.0

	COMMUNITY AREA NAME	CENSUS BLOCK	BUILDING TYPE	BUILDING_SUBTYPE	THERM JANUARY 2010	THERM FEBRUARY 2010	THERM MARCH 2010	THERM APRIL 2010	THERM MAY 2010	THERM JUNE 2010	THERM JULY 2010	THERM AUGUST 2010	THERM SEPTEMBER 2010	THERM OCTOBER 2010	THERM NOVEMBER 2010	THERM DECEMBER 2010
38278	Loop	1.703132e+14	Commercial	Commercial	3442.0	2174.0	2078.0	555.0	519.0	523.0	481.0	543.0	515.0	473.0	768.0	1941.0
38421	Loop	1.703132e+14	Commercial	Municipal	547.0	4673.0	2140.0	1982.0	1712.0	1687.0	1977.0	1873.0	1779.0	2129.0	2126.0	2720.0
38441	Loop	1.703132e+14	Commercial	Commercial	26533.0	22582.0	19806.0	16489.0	10517.0	6900.0	16185.0	12766.0	12725.0	14340.0	16231.0	25151.0
38442	Loop	1.703132e+14	Commercial	Municipal	37793.0	33336.0	23741.0	7320.0	4528.0	1217.0	1173.0	1146.0	898.0	4245.0	18462.0	35026.0
38443	Loop	1.703132e+14	Commercial	Commercial	33314.0	22873.0	7441.0	3285.0	2171.0	396.0	474.0	311.0	576.0	3903.0	8656.0	28522.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
38590	Loop	1.703184e+14	Commercial	Commercial	61037.0	64632.0	42911.0	21841.0	13181.0	6048.0	5479.0	6179.0	5863.0	8865.0	22988.0	60728.0
38591	Loop	1.703184e+14	Commercial	Commercial	30412.0	30380.0	23747.0	5349.0	152.0	258.0	108.0	93.0	287.0	603.0	3774.0	48667.0
38592	Loop	1.703184e+14	Commercial	Commercial	12883.0	11253.0	10453.0	4450.0	1684.0	103.0	124.0	140.0	101.0	1222.0	6670.0	11681.0
38593	Loop	1.703184e+14	Commercial	Commercial	25955.0	20581.0	13942.0	8535.0	6481.0	3076.0	2967.0	1428.0	4041.0	7933.0	13596.0	25341.0
38595	Loop	1.703184e+14	Commercial	Commercial	5362.0	7580.0	3591.0	2828.0	1361.0	679.0	709.0	687.0	641.0	1037.0	2247.0	4626.0

	COMMUNITY AREA NAME	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
1	Ashburn	916.750000	967.625000	526.750000	535.500000	314.750000	534.125000	570.750000	348.375000	419.625000	692.500000	1971.750000	2459.500000
5	Austin	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	273.000000	206.333333	138.666667	46.000000	0.666667
7	Austin	490.000000	441.666667	98.000000	130.333333	122.000000	734.666667	781.666667	677.333333	306.666667	195.333333	235.000000	261.666667
8	Austin	410.166667	814.666667	482.166667	456.166667	391.666667	506.166667	645.666667	810.166667	863.333333	497.333333	439.166667	599.500000
9	Austin	0.000000	0.000000	0.000000	0.000000	0.000000	170.333333	301.333333	606.000000	656.000000	246.000000	150.000000	735.666667
...	...	...	...	...	...	...	...	...	...	...	...	...	...
67046	Woodlawn	450.833333	219.666667	263.666667	244.166667	249.000000	498.333333	408.166667	391.833333	202.166667	362.333333	481.333333	837.500000
67047	Woodlawn	111.666667	195.555556	169.000000	203.555556	252.444444	262.333333	335.333333	336.666667	320.666667	425.888889	698.888889	1352.111111
67048	Woodlawn	509.571429	433.000000	368.857143	327.857143	1128.857143	712.428571	824.714286	570.857143	435.714286	443.285714	554.285714	669.142857
67049	Woodlawn	172.571429	150.714286	144.000000	158.428571	227.285714	195.285714	224.142857	221.571429	196.571429	176.571429	301.142857	361.285714
67050	Woodlawn	226.416667	254.750000	224.583333	316.083333	353.083333	448.583333	462.000000	577.416667	440.000000	497.583333	582.166667	428.666667

	COMMUNITY AREA NAME	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
0	Albany Park	588.960230	620.076751	599.319381	608.100158	787.060110	1062.593329	1124.960334	902.121725	681.448159	682.711545	910.421071	975.996995
1	Archer Heights	4060.960537	4162.342241	3963.701645	3882.010044	4056.491226	4694.134467	4813.099926	4811.352907	4152.028865	3892.331308	4081.929164	4534.551003
2	Armour Square	1742.539516	1624.745099	1512.367518	1574.774637	1571.659308	1762.314045	2198.402000	2139.742148	1774.726173	1223.895629	1602.952279	2310.849074
3	Ashburn	780.696213	745.711491	704.729112	717.169818	960.698133	1346.862270	1441.881978	1121.033684	808.984312	775.900670	1065.874121	1191.412444
4	Auburn Gresham	586.611070	590.521216	581.013643	573.039983	706.920065	921.194964	971.684396	820.333114	636.278627	667.405742	888.332942	1013.297035
...	...	...	...	...	...	...	...	...	...	...	...	...	...
72	West Lawn	1599.801869	1761.060292	1676.913074	1675.510384	2034.169928	2506.009592	2512.312836	2422.593366	1906.256466	1680.336738	1806.449746	2138.588987
73	West Pullman	336.840340	505.430414	494.363589	517.450046	575.883072	692.737455	884.599784	846.449675	645.562179	586.981731	700.753697	921.851030
74	West Ridge	631.225589	633.842594	606.552057	679.627136	901.395931	1193.336081	1153.017542	936.103499	684.780656	712.970023	945.586482	948.583799
75	West Town	554.456614	601.668226	571.151123	570.600857	693.789572	896.494764	1049.318744	1041.675432	811.528727	688.255666	781.993540	956.159867
76	Woodlawn	7110.092941	6651.934074	4857.665927	4664.439473	4299.869910	5068.639403	4576.080946	5510.521041	6255.137064	5391.800905	5675.902200	8230.214213

	COMMUNITY AREA NAME	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
14	Chatham	548.736974	563.949036	546.049245	538.207488	668.152943	878.640122	938.656570	761.947701	605.753641	633.530900	836.728679	922.702208
19	East Garfield Park	450.959905	518.789042	528.272949	529.217821	573.993923	679.083028	825.308169	823.947782	735.957543	613.447287	729.301227	905.244509
20	East Side	418.123974	438.244319	401.374926	423.017437	617.825313	897.358812	937.610734	685.952468	495.418421	505.783496	712.306064	686.252539
23	Englewood	504.576513	525.370981	499.136593	480.369505	547.117476	675.751533	757.998640	730.072981	565.111700	590.747475	785.415277	941.913156
25	Fuller Park	485.376146	507.384843	517.840736	497.809375	550.832397	628.657527	769.049517	776.020786	585.774026	542.345548	686.766241	819.261537
28	Grand Boulevard	612.172127	602.778075	509.061568	493.853598	547.317564	662.937125	795.866038	809.540480	638.899086	582.338788	723.965683	981.641705
29	Greater Grand Crossing	540.691549	563.919851	538.917779	543.887588	634.444308	820.860347	869.311150	814.108248	634.695363	645.272314	827.015719	982.295835
56	Oakland	462.199800	473.069618	423.576991	413.032848	426.406716	530.717199	676.657687	671.607153	533.147311	420.315195	576.774318	801.655847
62	South Chicago	447.606205	466.036084	462.377413	444.617988	530.067131	693.512976	748.631958	649.540879	510.929136	546.391435	724.367365	856.924134
65	South Shore	486.540287	498.248891	488.881648	486.547177	542.520038	726.045306	796.639112	713.272992	561.745431	573.613358	729.811821	876.291037
70	West Englewood	618.129398	594.753803	563.928640	544.140699	626.366991	785.657575	848.069335	796.693279	618.995212	650.060245	825.636205	964.519868
71	West Garfield Park	518.252198	598.153529	600.395014	585.378133	668.000861	797.586222	936.586922	904.084362	746.388751	636.364346	794.316026	985.578056
73	West Pullman	336.840340	505.430414	494.363589	517.450046	575.883072	692.737455	884.599784	846.449675	645.562179	586.981731	700.753697	921.851030

	COMMUNITY AREA NAME	THERM JANUARY 2010	THERM FEBRUARY 2010	THERM MARCH 2010	THERM APRIL 2010	THERM MAY 2010	THERM JUNE 2010	THERM JULY 2010	THERM AUGUST 2010	THERM SEPTEMBER 2010	THERM OCTOBER 2010	THERM NOVEMBER 2010	THERM DECEMBER 2010	GAS ACCOUNTS
0	Archer Heights	2326.0	2131.0	1400.0	620.0	502.0	224.0	222.0	187.0	197.0	252.0	744.0	2112.0	11.0
6	Austin	3041.0	2680.0	1151.0	373.0	124.0	26.0	29.0	25.0	49.0	177.0	670.0	3895.0	3.0
11	Austin	910.0	738.0	632.0	448.0	254.0	113.0	42.0	29.0	33.0	36.0	166.0	633.0	3.0
13	Avondale	535.0	458.0	481.0	307.0	194.0	145.0	119.0	102.0	88.0	109.0	141.0	249.0	4.0
19	Avondale	285.0	285.0	244.0	129.0	73.0	46.0	39.0	27.0	39.0	29.0	50.0	144.0	3.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
67046	Woodlawn	2166.0	1681.0	1858.0	1172.0	708.0	360.0	72.0	67.0	77.0	185.0	623.0	1800.0	9.0
67047	Woodlawn	985.0	1152.0	1238.0	630.0	475.0	192.0	141.0	162.0	144.0	210.0	653.0	1744.0	8.0
67048	Woodlawn	2202.0	1874.0	1647.0	906.0	645.0	346.0	84.0	150.0	150.0	260.0	694.0	1335.0	5.0
67049	Woodlawn	95.0	11.0	47.0	9.0	45.0	18.0	22.0	9.0	17.0	11.0	18.0	13.0	5.0
67050	Woodlawn	2372.0	1787.0	1449.0	718.0	572.0	286.0	155.0	134.0	161.0	303.0	588.0	1469.0	13.0

	COMMUNITY AREA NAME	KWH JANUARY 2010	KWH FEBRUARY 2010	KWH MARCH 2010	KWH APRIL 2010	KWH MAY 2010	KWH JUNE 2010	KWH JULY 2010	KWH AUGUST 2010	KWH SEPTEMBER 2010	KWH OCTOBER 2010	KWH NOVEMBER 2010	KWH DECEMBER 2010
2	Armour Square	1742.539516	1624.745099	1512.367518	1574.774637	1571.659308	1762.314045	2198.402000	2139.742148	1774.726173	1223.895629	1602.952279	2310.849074
3	Ashburn	780.696213	745.711491	704.729112	717.169818	960.698133	1346.862270	1441.881978	1121.033684	808.984312	775.900670	1065.874121	1191.412444
5	Austin	881.393662	1143.496215	1142.900845	1061.795445	1256.132520	1484.143676	1639.071569	1672.210766	1369.155481	1197.265957	1405.209377	1630.487812
7	Avondale	825.904543	831.643551	816.281594	789.980707	950.883245	1282.302588	1387.953706	1167.133834	916.402788	865.263593	1107.338454	1239.955574
8	Belmont Cragin	1185.054781	1236.872753	1123.117644	1101.673213	1304.915316	1531.453974	1669.345291	1593.902853	1244.509159	1140.136696	1415.916247	1619.806319
9	Beverly	819.153565	843.576326	780.968340	847.318220	1133.758536	1584.180429	1546.303610	1147.048840	919.139609	861.511896	1283.501531	1187.529238
11	Brighton Park	962.527296	975.527314	893.291030	889.663229	1002.113841	1197.505350	1419.792567	1421.544570	1102.163523	948.540256	1092.090235	1247.987735
12	Burnside	1659.831583	1511.910669	1522.988559	1500.095776	1773.631201	2003.929350	2109.256792	1780.390617	1527.236072	1577.450641	1935.369574	1969.445154
16	Clearing	1006.018576	971.486217	908.137255	881.398771	1136.920999	1462.454409	1656.843804	1493.196089	1065.599291	960.189565	1175.699334	1349.713413
17	Douglas	804.330181	1028.136170	1072.324103	965.833775	1068.902056	1234.492284	1899.938112	1635.922526	1441.719261	1134.206334	1304.732489	1830.144795
21	Edgewater	787.956060	765.620076	757.785832	808.236524	1013.694592	1337.047880	1292.964239	1106.686453	898.508448	926.612126	1261.304178	1167.221496
24	Forest Glen	998.415636	1029.897147	930.546497	958.553462	1314.293749	1723.190553	1744.001857	1327.503928	987.360230	1031.827405	1403.344445	1414.124895
27	Garfield Ridge	1124.419351	1047.446808	925.297925	946.223852	1165.278231	1545.691703	1802.146002	1671.712309	1116.258401	984.279463	1191.441888	1421.864072
31	Hermosa	704.177132	718.416720	727.294078	692.283694	800.147358	1129.596609	1218.280040	1131.554421	841.572864	779.246675	928.565326	1153.127346
32	Humboldt Park	880.847058	904.685828	855.461803	815.365556	947.857923	1097.151848	1229.160301	1228.869095	933.532279	871.939696	1056.782991	1275.612207
33	Hyde Park	860.503676	891.304441	824.430828	858.400679	941.763228	1181.456530	1410.719011	1354.890042	957.197711	888.457646	1082.287230	1202.991025
34	Irving Park	787.311409	804.157295	760.386940	767.441701	958.748132	1292.566553	1361.615334	1182.035092	891.198331	852.952136	1081.173961	1229.086884
37	Lakeview	1010.291796	989.955995	986.119385	1007.950787	1289.291345	1712.388964	1787.969638	1551.799873	1166.202127	1086.955514	1294.617318	1535.295256
38	Lincoln Park	1430.381825	1381.343191	1312.933273	1323.007504	1620.290012	1988.384519	2155.631354	2079.804864	1650.778356	1474.125476	1718.213343	2043.712750
39	Lincoln Square	1289.470311	1197.343900	1154.690204	1226.145015	1493.415860	1777.453604	1811.905439	1484.892539	1199.797229	1202.759471	1566.712233	1739.853156
42	Lower West Side	2238.734970	2324.962051	2563.762259	2365.772965	2584.996838	2652.793231	3096.807976	2938.655204	2632.480765	2665.254262	2639.321548	2796.924253
43	McKinley Park	1290.580346	1284.954295	1226.908180	1166.402172	1315.028792	1508.648989	1830.942342	1750.085608	1496.456765	1330.582396	1418.612205	1661.874531
46	Mount Greenwood	1166.047926	1227.524962	1103.540173	1150.895012	1594.086521	1975.182248	2139.973906	1763.025987	1267.471487	1239.876325	1451.904435	1575.857425
50	New City	2158.034873	2128.007395	1979.528847	2041.684600	2209.034956	2346.285104	2517.115221	2498.114585	2225.133858	2066.290594	2194.410749	1934.417876
52	North Lawndale	769.911303	848.381662	857.489972	859.260080	920.269797	972.349859	1220.240197	1214.814409	1094.312193	932.652531	1120.387548	1251.061516
53	North Park	1597.105270	1550.847146	1466.172978	1585.892708	1825.844100	2264.106142	2234.551794	1983.246059	1590.103452	1602.898242	2052.842765	1996.183347
55	O'Hare	1471.117048	1338.369856	1301.571878	1302.013968	1724.254600	2069.873388	1940.255138	1641.615568	1356.539485	1371.032938	1845.774332	1758.726383
57	Portage Park	743.415492	754.489564	715.730941	732.006116	945.573452	1270.487066	1380.737304	1095.934655	796.496626	792.842362	1018.084277	1128.016885
58	Pullman	2100.235399	2647.225804	2651.496623	2748.754931	2934.033984	3181.305584	3166.381242	2884.105464	2583.234329	2593.965732	3128.337629	3009.082393
63	South Deering	1497.917369	1666.825278	1562.935376	1525.794170	1663.397625	2033.259323	2009.443261	1851.622628	1590.819372	1734.424423	2186.510849	2079.765557
64	South Lawndale	1083.949928	1159.269373	1137.217089	1135.061677	1246.770658	1424.521890	1813.676266	1823.113284	1440.972191	1247.026243	1321.758601	1485.961881
66	Uptown	822.961222	812.293937	776.984800	795.939722	1045.913039	1334.576922	1396.770914	1124.679495	853.774196	839.126860	1085.585506	1211.331702
69	West Elsdon	1050.124284	1075.118131	1051.426395	1081.294813	1321.606673	1844.828496	1990.117801	1849.691242	1267.647503	1130.864014	1210.435294	1396.508081
72	West Lawn	1599.801869	1761.060292	1676.913074	1675.510384	2034.169928	2506.009592	2512.312836	2422.593366	1906.256466	1680.336738	1806.449746	2138.588987

	COMMUNITY AREA NAME	THERM JANUARY 2010	THERM FEBRUARY 2010	THERM MARCH 2010	THERM APRIL 2010	THERM MAY 2010	THERM JUNE 2010	THERM JULY 2010	THERM AUGUST 2010	THERM SEPTEMBER 2010	THERM OCTOBER 2010	THERM NOVEMBER 2010	THERM DECEMBER 2010
0	Archer Heights	211.454545	193.727273	127.272727	56.363636	45.636364	20.363636	20.181818	17.000000	17.909091	22.909091	67.636364	192.000000
6	Austin	1013.666667	893.333333	383.666667	124.333333	41.333333	8.666667	9.666667	8.333333	16.333333	59.000000	223.333333	1298.333333
11	Austin	303.333333	246.000000	210.666667	149.333333	84.666667	37.666667	14.000000	9.666667	11.000000	12.000000	55.333333	211.000000
13	Avondale	133.750000	114.500000	120.250000	76.750000	48.500000	36.250000	29.750000	25.500000	22.000000	27.250000	35.250000	62.250000
19	Avondale	95.000000	95.000000	81.333333	43.000000	24.333333	15.333333	13.000000	9.000000	13.000000	9.666667	16.666667	48.000000
...	...	...	...	...	...	...	...	...	...	...	...	...	...
67046	Woodlawn	240.666667	186.777778	206.444444	130.222222	78.666667	40.000000	8.000000	7.444444	8.555556	20.555556	69.222222	200.000000
67047	Woodlawn	123.125000	144.000000	154.750000	78.750000	59.375000	24.000000	17.625000	20.250000	18.000000	26.250000	81.625000	218.000000
67048	Woodlawn	440.400000	374.800000	329.400000	181.200000	129.000000	69.200000	16.800000	30.000000	30.000000	52.000000	138.800000	267.000000
67049	Woodlawn	19.000000	2.200000	9.400000	1.800000	9.000000	3.600000	4.400000	1.800000	3.400000	2.200000	3.600000	2.600000
67050	Woodlawn	182.461538	137.461538	111.461538	55.230769	44.000000	22.000000	11.923077	10.307692	12.384615	23.307692	45.230769	113.000000

	COMMUNITY AREA NAME	THERM JANUARY 2010	THERM FEBRUARY 2010	THERM MARCH 2010	THERM APRIL 2010	THERM MAY 2010	THERM JUNE 2010	THERM JULY 2010	THERM AUGUST 2010	THERM SEPTEMBER 2010	THERM OCTOBER 2010	THERM NOVEMBER 2010	THERM DECEMBER 2010
0	Albany Park	284.552873	245.087784	200.396134	108.203023	70.542487	42.627449	34.757723	30.944222	32.728462	39.958613	78.510739	194.184422
1	Archer Heights	521.052922	401.826920	315.615721	168.591343	108.174522	76.405301	73.128319	66.208327	66.431437	85.968940	163.912070	411.780009
2	Armour Square	583.664740	451.322075	487.174976	267.113941	193.111138	123.958385	110.153765	108.135963	111.420355	190.671587	321.742738	590.266422
3	Ashburn	287.763035	247.270869	235.673138	127.214171	65.933325	46.456245	28.085409	28.700234	26.699497	34.093083	63.663270	157.917898
4	Auburn Gresham	323.149316	298.781050	232.171357	126.028303	77.288771	42.824096	30.284555	29.387503	31.015608	50.249806	111.352088	274.396757
...	...	...	...	...	...	...	...	...	...	...	...	...	...
72	West Lawn	304.764895	266.360117	189.982854	107.564402	69.055657	45.487462	41.903536	41.695916	40.238220	49.723047	96.189277	247.669317
73	West Pullman	269.176338	251.701739	182.500750	97.649137	63.180883	28.492159	23.252768	22.823583	22.590791	38.725627	91.067161	228.711587
74	West Ridge	340.371726	278.831150	224.815387	133.599109	92.930962	64.734452	57.671696	51.438850	45.241085	51.940878	107.929262	247.420668
75	West Town	214.904567	183.123474	159.826261	78.785008	46.839160	30.773432	23.777846	21.285254	22.968842	28.651958	53.462065	148.915897
76	Woodlawn	371.700188	319.064021	285.381080	160.449406	100.195986	57.049176	39.392705	30.133532	41.408589	60.148513	118.230107	246.791872

	COMMUNITY AREA NAME	THERM JANUARY 2010	THERM FEBRUARY 2010	THERM MARCH 2010	THERM APRIL 2010	THERM MAY 2010	THERM JUNE 2010	THERM JULY 2010	THERM AUGUST 2010	THERM SEPTEMBER 2010	THERM OCTOBER 2010	THERM NOVEMBER 2010	THERM DECEMBER 2010
1	Archer Heights	521.052922	401.826920	315.615721	168.591343	108.174522	76.405301	73.128319	66.208327	66.431437	85.968940	163.912070	411.780009
2	Armour Square	583.664740	451.322075	487.174976	267.113941	193.111138	123.958385	110.153765	108.135963	111.420355	190.671587	321.742738	590.266422
5	Austin	375.527418	340.070165	256.766228	147.349960	101.834839	64.277464	59.122924	56.433242	54.945450	69.581625	143.622471	315.235657
11	Brighton Park	329.036687	270.078502	239.283083	132.245368	91.166026	68.551441	51.058345	52.244289	62.327791	64.676049	104.926773	232.598242
21	Edgewater	658.917630	570.229261	416.664880	231.148473	156.636148	79.454516	73.251229	71.783102	81.109346	154.394763	303.713912	625.323361
31	Hermosa	413.722371	368.280032	286.225292	183.922083	133.238923	110.972275	100.823497	105.861043	109.505919	138.939544	202.431431	382.412413
37	Lakeview	437.751977	382.584217	295.267978	166.219545	109.859130	67.378199	57.978375	56.030234	60.159221	92.558238	169.443953	352.299955
42	Lower West Side	443.226312	359.981507	382.984658	216.159551	142.063576	128.130953	111.958657	112.407813	98.313283	113.654632	172.630533	325.808930
43	McKinley Park	323.566614	280.694309	276.185830	140.544196	101.248356	80.469367	73.856148	57.956081	62.759469	75.724301	112.027159	234.817817
49	Near West Side	671.089125	541.241674	473.597290	257.104121	183.030549	143.469029	134.390934	135.583177	117.382970	152.604912	264.358699	555.532796
50	New City	475.741516	426.900098	319.378215	161.227358	118.385827	93.777355	86.933380	85.099924	84.569976	117.744776	227.770992	447.706175
53	North Park	411.453397	337.025392	269.959717	155.258982	94.467697	61.219057	58.709252	50.499208	47.940140	69.290616	135.998739	292.478673
65	South Shore	384.221957	313.447754	280.513328	169.539498	112.710766	73.521336	73.531265	54.686490	60.867523	77.075018	139.752888	267.479329
66	Uptown	456.350989	393.458826	323.010425	190.977445	122.452670	76.596368	64.350477	59.349936	60.648616	102.107306	179.623707	365.326735

	COMMUNITY AREA NAME	THERM JANUARY 2010	THERM FEBRUARY 2010	THERM MARCH 2010	THERM APRIL 2010	THERM MAY 2010	THERM JUNE 2010	THERM JULY 2010	THERM AUGUST 2010	THERM SEPTEMBER 2010	THERM OCTOBER 2010	THERM NOVEMBER 2010	THERM DECEMBER 2010
20	East Side	234.096488	205.636718	200.159872	104.559541	60.048371	40.656037	24.424548	21.883132	22.843186	27.464492	52.064187	144.921484
30	Hegewisch	256.971420	209.963993	185.351310	99.199959	57.086837	33.176131	25.057912	21.692122	23.137106	31.599408	69.430850	158.439756
35	Jefferson Park	267.822697	239.742414	186.320303	110.622652	58.567879	37.180809	27.102508	24.842575	25.624770	36.801825	82.166749	190.918519
40	Logan Square	237.602956	220.901166	200.925168	95.981718	62.083632	42.810736	34.941626	32.565628	33.249863	46.279649	87.237568	204.288151
51	North Center	251.140221	211.428602	175.547574	92.889053	56.446202	32.115566	25.250018	20.648477	20.614367	32.650028	67.422616	164.334048
75	West Town	214.904567	183.123474	159.826261	78.785008	46.839160	30.773432	23.777846	21.285254	22.968842	28.651958	53.462065	148.915897

	COMMUNITY AREA NAME	THERM JANUARY 2010	THERM FEBRUARY 2010	THERM MARCH 2010	THERM APRIL 2010	THERM MAY 2010	THERM JUNE 2010	THERM JULY 2010	THERM AUGUST 2010	THERM SEPTEMBER 2010	THERM OCTOBER 2010	THERM NOVEMBER 2010	THERM DECEMBER 2010
41	Loop	10605.645853	8416.081631	6183.463934	3435.484118	2396.747512	1592.728842	1520.397669	1634.239188	1613.739611	2495.810977	5013.781157	9956.751008
47	Near North Side	2561.680909	2119.106513	1683.767077	1062.919003	824.563958	593.747481	700.338868	737.632438	570.851880	823.457823	1344.985954	2320.957193
48	Near South Side	2039.334959	1723.952502	1085.831066	719.351076	476.540802	226.049091	200.692879	200.828904	231.522441	491.562926	950.021815	1913.890225
59	Riverdale	1872.426728	1824.674491	1665.893264	1441.338703	935.926139	894.621183	1076.188708	1065.118085	1060.908624	1279.218855	1444.840866	1990.056394

Consumo de energía en Chicago en 2010¶

Contenido¶

Exploración de los datos¶

Estudio de datos¶

Outliers¶

Energía¶

Gas¶

Comunidades que existen¶

Ensayos¶

Clustering de energía por tipo de zonas agrupando por ciudades¶

Comercial¶

Residencial¶

Industrial¶

Clustering de gas por tipo de zonas agrupando por ciudades¶

Comercial¶

Residencial¶

Industrial¶

Clustering de energía para cada comunidad observando patrones para cada bloque¶

Comercial¶

Residencial¶

Industrial¶

Clustering de gas para cada comunidad observando patrones en cada bloque¶

Comercial¶

Industrial¶

Residencial¶

Clustering de energia por media consumida en cada contador por comunidad¶

Separación por cuartiles¶

Clustering de gas por media consumida en cada contador por comunidad¶

Separación por cuartiles¶

Clustering por edad del edificio¶

Energía¶

Gas¶

Clustering por dimensión del hogar¶

Energia¶

Gas¶

Clustering por ocupación total¶

Energía¶

Gas¶

Clustering por porcentaje de casas en renta¶

Energía¶

Gas¶